• EPILEPSY
• EPIDEMIOLOGY
• ANTICONVULSANTS
• OBSTETRICS
• PHARMACOLOGY

We undertook a population-based study to assess the comparative risk of spontaneous abortions, terminations of pregnancy (TOPs), major birth defects (MBDs), preterm births and small for gestational age (SGA) infants following intrauterine antiepileptic drug (AED) exposure in the Emilia Romagna region, Northern Italy (4 million inhabitants) over a 3-year period (2009–2011). The study was approved by the Local Ethical Committee and authorised by the Hospital Management Executive.

Data were obtained from official regional registries: Certificates of Delivery Assistance (CedAP) for deliveries (>99% coverage of live births and stillbirths) and Hospital Discharge Cards collecting the International Classification of Diseases (ICD) codes of all the discharge diagnoses, for abortions . AED exposure was based on reimbursed prescription databases as AEDs are supplied and reimbursed by medical prescription in Italy; all drugs registered as AEDs and redeemed with ATC code N03 were considered. MBDs were taken from the Registry of Congenital Malformations (IMER) recording MBDs detected by paediatricians in stillbirths and in live births during the first week of the infant’s life (coverage >95% of births). Each recording from the first three registries contains a unique anonymous code identifying each citizen and linking information among registries. Recordings from the IMER are routinely linked to the CeDAP and consequently to the others, using an additional procedure (correct linkage >98% of the sample). All data were anonymised at the regional source of data to protect patient privacy according to Italian legislation. For the deliveries cohort, the exposure period was based on the date of delivery and gestational age. For the abortion cohort, the trimester preceding the event was fixed by consensus among the authors as the AED exposure period in pregnancy, gestational age being untraceable. SGA was defined as birth weight below the 10th percentile of the gender-specific birth weight for gestational age reference curves. MBDs were coded using a British Paediatric Association modification of the WHO ICD-9 system (BPA, 1979). Sample size was calculated to identify a minimal detectable risk of MBD of 1.9% (beta error 20% and alpha error 5%) in the whole population for exposure to any AED.

Among 145 243 pregnancies (111 284 deliveries, 112 845 live births and 279 stillbirths, 16 408 spontaneous abortions and 17 551 TOP), 21 (0.014%) were exposed to gabapentin (GBP), 30 (0.02%) to pregabalin (PGB) and 1 was exposed to both drugs. Among pregnancies exposed to GBP only, two ended in spontaneous abortion (9.5% vs 11.3% in unexposed—OR=0.83, CI 95%=0.09 to 3.43; p=0.6), and eight in TOP (38% vs 12% in unexposed—OR=4.51, CI 95%=1.62 to 11.75; p=0.0021). Among the 11 newborns exposed to GBP only, 6 were born preterm (54.5% vs 14%—OR=7.37, CI 95%=1.87 to 30.54; p=0.0018) and 4 were SGA (36.3% vs 10%—OR=5.14, CI 95%=1.10 to 20.23; p=0.018). Among the nine newborns exposed to GBP during the first trimester, two had an MBD, an isolated ventricular septal defect (VSD) in both cases, while one of the two was also SGA. Among the pregnancies exposed to PGB only, seven ended in spontaneous abortion (23.3% vs 11.3% in unexposed—OR=2.39, CI 95%=0.87 to 5.75; p=0.07), and seven in TOP (23.3% vs 12%—OR=2.23, CI 95%=0.81 to 5.37;p=0.08). Among the 16 newborns exposed to PGB only, 4 were born preterm (25% vs 14%—OR=2.05, CI 95%=0.48 to 6.76; p=0.27) and 2 were SGA (12.5% vs 10%—OR=1.29, CI 95%=0.14 to 5.60; p=0.67). Among the 13 newborns exposed to PGB during the first trimester, 1 had an MBD, namely a VSD, and was also SGA. The baby exposed to both drugs was SGA.

The few literature data currently available on GBP human teratogenicity seem to point to a low risk of MBDs.1 However, Morrow et al 2 found that the only malformation out of 11 exposures was a VSD, while Molgaard-Nielsen and Hviid3 reported that the only malformation out of 59 exposures was an unspecified ‘congenital heart disease’. A slightly higher risk of preterm birth and low birth weight after intrauterine exposure to GBP was reported by Fujii et al.1 In this cohort, however, no increased risk of SGA was found.1 Data on the teratogenicity of PGB are even more scant but a recent population study found an increased risk of birth defects, whereas the rate of spontaneous abortions was not increased.4

GBP and PGB are gamma-aminobutyric acid analogues that share several pharmacodynamic and pharmacokinetic features. Both have shown high affinity for the α 2d calcium channel subunit expressed in brain, spinal cord, skeletal and cardiac muscles. In particular, the α 2d-1 and α 2d-2 subtypes have a set of three arginine residues that are required for GBP and PGB binding. The α 2d-1 subtype plays a key role in muscle and neuronal tissue development, namely myoblastic attachment, migration and spreading. Impaired function of this subtype can therefore cause dysregulation of muscle formation.5 The incomplete heart development found in our infants could be explained by exposure to GBP and PGB in the early stages of pregnancy.

The teratogenic effect could be triggered by the L-amino acid transporter (LAT) present in the intestine, brain and placenta. LAT facilitates gut absorption of both PGB and GBP and their sequential passage across the blood–brain barrier. The same mechanism could also facilitate drug transfer across the placenta to the fetus.

GBP and PGB are both widely and increasingly used to treat a number of conditions other than epilepsy, including pain and anxiety. Our population-based study was based on prescription data and medication adherence could not be verified. Moreover, our study was underpowered to provide information on the risks associated with single antiepileptic drugs and larger studies are needed. However, our findings should raise concerns over the use of these drugs in women of childbearing potential.